The Indian construction industry, which is presently executing over 240 km of tunnel length particularly for metro rail and water infrastructure projects using tunnel-boring machines (TBMs), is steadily adopting technologies such as tunnel seismic prediction, ground penetrating radar, AI-enabled analytics and real-time geotechnical monitoring to improve geological certainty before excavation.Experts we reached out to pinpointed managing uncertainty in subsurface conditions as the most persistent challenge in tunnelling. Especially, in the Himalayan region and other geologically complex terrains, according to Rishabh Gupta, Design Manager, RITES.Considering that every tunnelling project involves detailed geological investigations during planning, why do uncertainties arise and how are leading construction companies overcoming blind spots?Point-specific investigation drawbacksEven the most comprehensive geological investigations provide only point-specific information and cannot cover the continuously changing ground conditions a TBM advances through, explains Yogesh Srivastava, VP & Head - Fleets & Resources, Tata Projects.Consequently, tunnelling frequently encounters unforeseen ground conditions such as shear zones, fault gouge, excessive groundwater ingress, squeezing ground or highly fractured rock masses.“Mismatches between the predicted and actual geology are a real project risk,” according to Srivastava, with the potential to directly impact excavation rates, cutter life, machine availability and, ultimately, project schedules. Fortunately, ways exist to mitigate tunnelling uncertainty and avoid the associated construction delays, increased costs, redesign and, in extreme cases, safety incidents.From static to adaptive designThe solution to tunnelling uncertainty, according to Gupta, based on his railway and metro tunnel project experience, is shifting from a static design philosophy to a data-driven and adaptive approach. “During rehabilitation studies for Kolkata Metro Line 1 and the Jammu-Katra rail link, while supporting tunnel design and slope stability works for railway projects in challenging terrains, we integrated advanced site investigations with non-destructive testing, hydrogeological assessments and three-dimensional geotechnical modelling.”Technology enabled his team to better understand the in-situ behaviour of the ground and optimise rehabilitation and support measures based on actual field conditions rather than conservative assumptions.Gagandeep Kansal, COO, Amberg Engineering India, proposes reducing the uncertainty of ground behaviour prediction by combining conventional geotech investigations with advanced techniques like geophysical, electrical resistivity tomography (ERT), seismic refraction tomography (SRT), transmission electron microscopy (TEM), etc. and AI/ML-based geological prediction.Azhar Jamil, Sr. Manager (Geotechnical and Underground Expert), DMR Engineering, proposes combining digital geological mapping, LiDAR/photogrammetry, probe drilling, real-time monitoring and data analytics with conventional Q/RMR classification and engineering judgement. These technologies enable faster decisions and optimisation of excavation and support based on the actual ground conditions.For example, on the Lower Solu Hydropower Project, Nepal, implementing design solutions through low-to-high squeezing conditions, continuous geological assessment and timely support decisions proved critical and helped commission a challenging project.Further, Kansal points out, “Real-time monitoring and instrumentation and IoT sensors that can help monitor TBM operational parameters like torque, thrust, penetration rate, vibration, water pressure, gas leaks, etc, help in safe tunnel advancement during mining. Using advanced geophysical techniques such as tunnel seismic prediction ahead of the tunnel face during tunnelling can provide a fair assessment of the anomalies in ground behaviour leading to better preparedness in advance.”Tunnel seismic predictionSrivastava proposes switching from reactive to predictive tunnelling, which implies knowing what lies ahead before getting there, to reduce uncertainty.“Technologies such as tunnel seismic prediction use seismic wave analysis to map geological conditions 25 to 100 m ahead of the tunnel face, enabling engineers to anticipate rock transitions, fault zones, fractured strata and water-bearing zones before excavation reaches them,” explains Srivastava. “In response, project teams can optimise TBM operating parameters, proactively plan cutter replacement and maintenance, and minimise schedule uncertainty before it affects project execution.”Tata Projects faced this challenge during the Chennai Metro underground project when a comparison between the geological investigation report and the actual geology encountered during tunnelling showed deviations across sampled tunnel sections in two-thirds of the locations. In one stretch, the proportion of harder Grade II-III rock increased from an expected 20 per cent to an actual 65 per cent.Srivastava’s teams applied higher cutter head torque but saw lower penetration rates and reduced advance speeds, prompting probe drilling and cutter head interventions to validate the geology and replace worn cutting tools. While this approach ensured safe execution, it remained reactive.“Each cutter head intervention typically resulted in seven to 10 days of downtime, impacting 35-50 rings of planned progress at an average advance rate of five rings per day, while also increasing cutting tool consumption, equipment idling and project overheads,” says Srivastava.On large underground projects, avoiding a single unplanned cutter head intervention can protect 5-10 per cent of the tunnelling schedule, he estimates.Future perfectWith a tunnelling pipeline comprising over 500 ongoing tunnel packages spanning more than 1,100 km, and another 1,350+ packages covering over 1,800 km in the planning or bidding stage, the stage is set for a tech invasion in tunnelling.Digital twins can help to continuously compare the predicted versus actual ground behaviour, according to Kansal. In future, Gupta expects the integration of digital twins, BIM, AI and real-time instrumentation to prove transformative for tunnelling.“Continuous monitoring through automated total stations, convergence monitoring, piezometers, inclinometers and fibre-optic sensing can feed live data into digital models, allowing engineers to predict ground behaviour, identify anomalies at an early stage, and modify excavation or support systems proactively,” says Gupta. “AI-based analytics can further enhance decision-making by identifying patterns from historical and real-time data that may not be immediately evident through conventional analysis.”The future of tunnelling is not technology replacing experience – it is technology enabling experienced engineers to make faster, safer and more informed underground decisions, says Jamil.Not only stronger excavation equipment but also smarter engineering decisions driven by continuous data acquisition, predictive analytics and digital integration are vital to significantly improve safety, reduce project risks, optimise costs and enhance the resilience of underground infrastructure.Tech-based open trenchingTechnology that converts uncertainty into actionable intelligence and converts a reactive approach to a proactive one helps reduce tunnelling risks and leads to better project management.In that context, “Sing ground penetrating radar in open to micro-trenching and horizontal directional drilling helps identify subsurface utilities and reduce construction risks,” says Azhar Jamil, Sr. Manager (Geotechnical and Underground Expert), DMR Engineering.Open trenching involves multiple activities: Utility diversion, shoring, excavation, dewatering, bedding and pipe laying or cut and cover tunnel construction and backfilling. In all this, Gagandeep Kansal, COO, Amberg Engineering India, says, “Productivity is often affected by improper utility mapping and clashes leading to rework, equipment waiting/idling time leading to loss of time and increase in quantities, throwing the project out of schedule.”So, Kansal recommends advanced GPRS and GIS technology to properly identify and map existing utilities and integrate with BIM models to avoid clashes, IoT-based sensors for equipment to measure utilisation and idle time, and AI-based progress monitoring by comparing site reality with the BIM model/schedule.